Monday, June 25, 2018

But Wait a Sec . . . Is that really the Best Way?

 A Less-than-Ideal Example

After I wrote the previous blog post, I had a good think about a couple of things. One was that my situation with Anovisions provides, occasionally, less-than-ideal entry points into studies.I often enter a study long after many of the critical first decisions about data gathering, data type, and even research questions have been made by the people who have decided to hire a statistician. My methods for getting my head around a study, which I have started to describe in the previous blog post, are therefore rather ad hoc and perhaps not the best methods for researchers in environments where they have more control over the methods.

At this point in the conversation, therefore, I'm going to go in a couple of different directions. Let's have a look at a process that is more or less ideal for a data explorer, and after that, let's go back to how to manage when you are required to be the statistics expert entering in the middle of an exploration that has already been begun by people who are savvy in their businesses but not necessarily trained as data scientists or statisticians.

An Ideal Example

In this example, you have just been handed data and you have been asked to see what you can learn from it. It's a data scientist's dream: you can follow a process that allows you to find whatever the data has to offer.

Choose Your Tools

Roger Peng, PhD of Johns Hopkins is an articulate
proponent of the R Programming Language. 
We have our pick of tools, so let's use R to do the analysis. It's powerful, flexible, free, open-source data analysis software and I love it. So does Roger Peng, PhD of Johns Hopkins University, and he says why in this video that introduces a really fun course on R programming that you might want to take some day. I did and it was excellent.

Installing R and RStudio

To learn how to install R and RStudio on your own computer so you can follow along with the data analysis step by step, please see this video.
To learn how to install R and RStudio on a Mac, see this video. Dr. Peng talks fast, so you may need to pause frequently as you follow along on your computer.

We will continue in this vein for our next few posts.

This time I'll let somebody else do the talking.

Every now and again you come across an answer to a question that is just so clear that you want to bookmark it so that you can find it again if something like the original question you were googling comes up again. That just happened to me. I was googling "How are linear regression models, ANOVAs, and ANCOVAs related?" because it is a long time since I have had to do those equations by hand and all I remembered was that I learned them together because they relied on the same basic formulae. What those formulae were, I had completely forgotten.

Thanks, SAS, SPSS, Python, and R, for doing all the work for me and letting my brain atrophy.

Here is an answer to my question that is so good that I'm giving it space on my blog. Thanks, Karen Grace-Martin, for tapping this out so the rest of us could have the benefit of your explanation.